Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 16 Feb 2008 13:33:24 GMT
From:      rwatson@FreeBSD.org
To:        sam@inf.enst.fr, rwatson@FreeBSD.org, freebsd-bugs@FreeBSD.org
Subject:   Re: bin/24271: [patch] dumpon(8) should check its argument more
Message-ID:  <200802161333.m1GDXOaJ010716@freefall.freebsd.org>

next in thread | raw e-mail | index | archive | help
Synopsis: [patch] dumpon(8) should check its argument more

State-Changed-From-To: open->closed
State-Changed-By: rwatson
State-Changed-When: Sat Feb 16 13:27:29 UTC 2008
State-Changed-Why: 
After some discussion on #freebsd-bugbusters, it sounds like we should close
this PR on the basis that UNIX likes to let you shoot your feet.  If we were
designing swap/dump behavior from scratch, my preference would be to have a
partition header in much the same way we have a superblock for file systems
so that partition selection could be validated directly.  However, in the 
ame way that we don't validate partions for swap, we don't validate for
dump, as dump and swap data may contain arbitrary contents.

While it's not something we appear to heavily support in fstab anymore, I
actually like the ability to dump to a separate partition or location from
swap -- swap has a tendency to get overwritten for large file system checks
(etc), whereas a dedicated dump partition isn't.  Likewise, on very small
systems where textdumps are used and swapping isn't enabled, that
distinction is also useful.


http://www.freebsd.org/cgi/query-pr.cgi?pr=24271



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200802161333.m1GDXOaJ010716>